Why 3'-UTR length is 1 or 0?

Question

[yancychy]

Hi, I want to get the 5'-UTR and 3'-UTR coodinates from the annotation file. I downloaded the "All GENCODE VM24" bed file from UCSC genome browser. The format of the bed file is following:

bin                 name   chrom  strand   txStart      txEnd            
1023 ENSMUST00000000090.7    chr9      +     57521278   57532426    

cdsStart       cdsEnd         exonCount 
57521327       57531782         5

exonStarts                                        exonEnds
57521278,57528955,57530240,57531668,57532281,     57521415,57529072,57530362,57531792,57532426,       

score  name2    cdsStartStat   cdsEndStat  exonFrames 
0      Cox5a         cmpl      cmpl       0,1,1,0,-1,

For 5'UTR, its length is cdsStart - txStart = 57521327 - 57521278 = 49, from (57521278, 57521327 ]

For 3'UTR, its length is txEnd - cdsEnd = 57532426 - 57531782 = 644, from (57531782, 57532426 ]

However, the length of all exons (the trs) is only 645.
https://useast.ensembl.org/Mus_musculus/Transcript/Sequence_cDNA?db=core;g=ENSMUSG00000000088;r=9:57521279-57532426;t=ENSMUST00000000090

Is there anything wrong?

In addition, I find the 5'UTR and 3'UTR length are 1 for some transcript. Is it reasonable? In the link A: Easy Way To Get 3' Utr Lengths Of A List Of Genes, the 3'UTR length of OR4F5 is 0 and 1.

5utr ensembl_transcript_id  
A    ENSMUST00000054837

3utr ensembl_transcript_id   
G    ENSMUST00000073261

Answer 1

[i.sudbery]

The end of the coding sequence is not in the last exon, but the penultimate exon -

the cdsEnd is 57,531,782 and the start of the last exon is 57,532,281. Thus your calculation for the length of the UTR is including the last intron in the sample.

I would guess in terms of the 1bp or 0bp UTRs, that they are simply unannotated.

EDIT: That first one is a non-coding RNA. It shouldn't have any UTR (or rather its all UTR, so there is no 5' or 3' UTR). The second one is protein coding. Its a Histone gene. Histones have odd termination machinary (they don't use polyA signals), but as far as I was aware, they do have UTRs normally.

[yancychy]

Thanks very much. The real 3'UTR length is 57531792 - 57531782 + (57532426 - 57532281) = 155. Thanks for the explaination of non-coding RNA and the UTR length. It's very clear.